BEGIN:VCALENDAR
PRODID:-//Columba Systems Ltd//NONSGML CPNG/SpringViewer/ICal Output/3.3-
 M3//EN
VERSION:2.0
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20260224T144758Z
DTSTART:20260311T110000Z
DTEND:20260311T120000Z
SUMMARY:AI-Fun & ELLIS Invited Speaker Series | Nikolay Malkin
UID:{http://www.columbasystems.com/customers/uom/gpp/eventid/}x1en-mkdxxx
 68-nm88s3
DESCRIPTION:For March's AI-Fun and ELLIS invited speaker series\, we will
  have Nikolay Malkin from the University of Edinburgh.\n\nTitle: Inferri
 ng stochastic dynamics without data: from diffusion samplers to discrete
  Schrödinger bridges\n\nAbstract: \nProbabilistic models that approximat
 e a distribution by transporting particles from a source distribution to
  the target following a learnt dynamics model have seen rapid developmen
 t and adoption in recent years: indeed\, diffusion models and continuous
  normalising flows show success in generative modelling for various doma
 ins. I will describe the less-known use of such dynamics-based models as
  variational families: fitting their parameters to sample distributions 
 from which no samples are available but an unnormalised target density c
 an be queried. This problem has many algorithmic faces\, with connection
 s to entropic reinforcement learning\, optimal transport\, stochastic co
 ntrol\, and sequential Monte Carlo. Our recent work has extended algorit
 hms for diffusion sampling to the discrete-space case and to learning br
 idge dynamics between two distributions without access to samples from b
 oth. Applications include sampling Boltzmann densities of molecular conf
 ormations\, inverse problems and conditional generation under pretrained
  generative model priors\, and accelerating particle-based algorithms fo
 r Bayesian inference (in the continuous case) and inference over probabi
 listic model structure (e.g.\, Bayesian program induction and symbolic r
 egression) and alignment of discrete-latent image generative models (in 
 the discrete case).\n\nBio:\nNikolay Malkin is a Chancellor's Fellow in 
 Informatics at the University of Edinburgh and a fellow of CIFAR's Learn
 ing in Machines and Brains programme. Their research focuses on algorith
 ms for probabilistic inference and Bayesian machine learning\, with appl
 ications in generative modelling\, neurosymbolic AI\, and machine reason
 ing. Within machine learning\, their work explores modelling of Bayesian
  posteriors over high-dimensional and structured variables\, induction a
 nd discovery of compositional structure in generative models\, and uncer
 tainty-aware reasoning in language and formal systems. Their work has fo
 und applications in pure and applied sciences\, including inverse imagin
 g\, remote sensing\, discovery of novel biological and chemical structur
 es\, and\, most recently\, robot control. Dr Malkin holds a PhD in mathe
 matics from Yale University (2021) and was previously a postdoctoral res
 earcher at Mila – Québec AI Institute in Montréal (2021 to 2024).\n\nIf 
 you are unable to attend in person\, please follow the ticketsource link
  provided to register and then check the registration confirmation for t
 he Teams link.
STATUS:TENTATIVE
TRANSP:TRANSPARENT
CLASS:PUBLIC
LOCATION:Lecture Theatre 1.3\, Kilburn Building\, Manchester
END:VEVENT
END:VCALENDAR